Feb. 7, 2007 First GLAST symposium 1

Measuring the PSF and the energy resolution with the

GLAST-LAT Calibration Unit

Ph. Bruel

on behalf of the beam test working group

Gamma-ray Large Gamma-ray Large Area Space Area Space TelescopeTelescope

Feb. 7, 2007 First GLAST symposium 2

Measuring the PSF and the energy resolution

• Data/Simulation agreement and the LAT analysis

• Beam tests at CERN (PS and SPS)– Tracker and PSF– Calorimeter and energy resolution

• Conclusions

Feb. 7, 2007 First GLAST symposium 3

Data/simulation agreement and analysis

• Event processing steps :– Trigger and On Board Filter– Reconstruction algorithms– Classification trees ( -> IRFs and Background rejection )

• Example 1 : PSF– Tracks, vertices are reconstructed– Using not only tkr variables, the classification tree gives :

• The best estimate of the incoming direction of the gamma• A quality variable used to select the best measured gammas

• Example 2 : Energy resolution– 3 energy reconstruction algorithms -> 3 estimates– Using not only cal variables, the classification tree gives :

• The best estimate of the energy of the gamma• A quality variable used to select the best measured gammas

• The calibration unit is different from the LAT– We can not apply directly and simply the classification tree analysis– But we can compare between data and MC

• the PSF of each direction estimate• The energy resolution of each energy estimate

Feb. 7, 2007 First GLAST symposium 4

Dealing with a huge phase space…

1700 runs, 94M events, 330 configurations (particle, energy, angle, impact position)

For PSF and energy resolution analysis : gammas (->2.5 GeV), electrons (5->280 GeV)

Feb. 7, 2007 First GLAST symposium 5

Mass simulation

• Simulation of the beam upstream the CU (Geant4)– Trigger/veto scintillators & cerenkov– Electron tagger at PS

• LAT simulation (GLEAM-Geant4)• Automatic generation of the run

configuration– Allowing the best comparison

• Efficient processing through the pipeline at SLAC

• Geant4 optimization– Simulation parameters– Physics lists

Feb. 7, 2007 First GLAST symposium 6

Tracker (work in progress…)

• 10-20% more hits in data than in MC• Effect almost independent of particle, E• Clusters are well reproduced by MC• More secondary particles / preshower ?

– Lower range cuts in Geant4 : no effect– Testing low energy processes– The calorimeter should say if there is

extra material upstream the CU

Hit multiplicity for gammas (0.5 to 2.5 GeV)

Hit multiplicity for 100 GeV electrons

Cluster-axis distance for 20 GeV electrons

Feb. 7, 2007 First GLAST symposium 7

Tracker (work in progress…)

• Preliminary PSF results for events with 1 vertex and 2 tracks shows a good agreement between data and the simulation

• This not yet the PSF from the IRFs (i.e after signal selection and background rejection)

<- 68%

PSF

95% ->

Feb. 7, 2007 First GLAST symposium 8

Calorimeter (work in progress…)

• Importance of quality/fiducial cuts• Raw energy : 10% more energy in

data for electrons at SPS

Feb. 7, 2007 First GLAST symposium 9

Calorimeter (work in progress…)

• The beam tests allow us to fully understand the electronics of the calorimeter– High rates– Cross-talk– Non linearity

• The 10% energy excess at high energy is mainly due to calibration issues– 2diodes x 2gains : 4 ranges from 1 MeV -> 70

GeV– Ground calibration : muon peak at 11 MeV– CERN data allow a complete check of our

procedure to extrapolate the energy scale from 11 MeV up to 70 GeV

Feb. 7, 2007 First GLAST symposium 10

Calorimeter (work in progress…)

• Very good agreement at PS energies• Thanks to the electron tagger, we can

measure the energy resolution from 50 MeV up to 2.5 GeV

• After naive calibration correction, very good agreement at SPS energies

280 GeV electrons at 30 deg

Feb. 7, 2007 First GLAST symposium 11

Conclusions

• A huge amount of good quality data to ensure that our simulation well reproduces the data through the LAT very large phase space

• Data reprocessing and mass simulation are fast and easy

• Ongoing tests to understand the current low-level disagreements

• The preliminary results show that PSF and energy resolution are well reproduced by the simulation